Photos are taken regularly throughout the day. At the end of the day a 4k time-lapse is generated and uploaded to YouTube.

Currently running at Nipomo Sunset Pi on YouTube.

• Add configuration to save data to an external drive.
  • Daily backup to a network drive.
• Use localized sunrise/sunset times to start and stop the time-lapse.
• Slow down the time-lapse around the sunset.
  • Make a separate sunset-only time-lapse without affecting the day-long time-lapse.
• Automatically add title and end screens
  • Uses a photo with a Ken Burns effect with the date and music information in the subtitle.
  • Make the title photo a picture of the daily sunset.
• Make the YouTube thumbnail be about 15 minutes before sunset.
• Automatically add royalty free music to the videos.
  • Would be cool to automatically scrape the YouTube audio library.
  • More likely I'll download a bunch of music, put it into a folder and cycle through them. (this is what I did)
• Add image stabilization to remove wind shaking.
• Make long term time-lapses of solar noon, and of each sunset showing how the sun moves across the horizon as the seasons pass.
  • Automatically make a monthly time-lapse.
  • Make a yearly time-lapse.
  • Make multi-year time-lapses.
• Automatically maintain monthly playlists on YouTube (August 2020, September 2020, etc.).
• Make the latest video always the featured video on the channel.
• Integrate with weather API to show daily high and low teperatures and daily rainfall.
• Add an AI voice over and shorten the title
• Add links to previous year's videos
• Automatically upload monthly time-lapses

• imagemagick - for resizing and cropping
• ffmpeg - for encoding video
• Google API Python Client - for uploading to YouTube

• Raspberry Pi 4 4-8GB (tested with a pi 3 b+ but was unable to encode higher than 1080p)
• Raspberry Pi HQ Camera
• Raspberry Pi HQ Camera Lens - Wide Angle
• Raspberry Pi PoE HAT
• Outdoor CCTV Camera Housing
• USB External Drive - it generates more than 5 GB of images a day and about 2 TB a year (assuming you want to save the raw images).

1. Setup a Raspberry Pi (preferably in headless mode, but not required)
2. Install an external hard drive and edit fstab so that it always mounts to the same location.
   • This may be optional if you don't care to save the raw data.
   • Carefully follow the instructions found here.
3. Install dependencies:

sudo apt install ffmpeg
sudo apt install imagemagick
sudo apt install python3-pip

# I like to have an alias that forces the python3 and pip3... so that's what I do:
echo 'alias python=python3' >> ~/.bash_aliases
echo 'alias pip=pip3' >> ~/.bash_aliases
source ~/.bashrc

# install a virtual environment to install all the YouTube specific stuff we need
cd
pip install virtualenv
python -m virtualenv youtube
source youtube/bin/activate
python -m pip install google-api-python-client
python -m pip install oauth2client
deactivate

4. clone this project:

cd
git clone https://github.com/kevinabrandon/sunsetpi.git

5. Download the NOAA Solar Calculations for the year
   • Go to the NOAA Solar Calculator and enter in your gps location
   • Click the "Create Sunrise/Sunset Tables for the Year" button
   • Using the mouse select the text of each table and copy and paste them into a spreadsheet program (I used google sheets).
   • Export each table as a csv file called:
     • YYYY-sunrise.csv
     • YYYY-sunset.csv
     • YYYY-solarnoon.csv
   • Make the YYYY the current year
   • Save the csv files into ~/sunsetpi/solar-tables/
6. setup crontab:

crontab -e

Add the following lines to the contab:

# trigger the camera every minute:
* * * * * . $HOME/sunsetpi/config.sh; $SUNSETPI_PATH/trigger_cam.sh
# trigger the camera on the 15, 30 and 45 seconds of each minute for the sunset portion:
* * * * * . $HOME/sunsetpi/config.sh; $SUNSETPI_PATH/trigger_cam.sh 15
* * * * * . $HOME/sunsetpi/config.sh; $SUNSETPI_PATH/trigger_cam.sh 30
* * * * * . $HOME/sunsetpi/config.sh; $SUNSETPI_PATH/trigger_cam.sh 45

# trigger the daily time-lapse creation at 9:15 pm every day:
15 21 * * * . $HOME/sunsetpi/config.sh; $SUNSETPI_PATH/make_daily_timelapse.sh

7. Setup your project on Google API Console
8. From the Google API Console create oauth2 credentials and put them in ~/sunsetpi/credentials.json
9. After the first day of photos manually run the make_daily_timelapse.sh so that you are able to follow the youtube authorization instructions in the console.
10. For now on the auth is saved and the make_daily_timelapse.sh can be run via cron each day.